Artificial intelligence has dominated enterprise conversations for more than a decade, but CES 2026 marked a clear inflection point: AI is no longer confined to dashboards, copilots, or back-office optimization. It is moving decisively into the physical world. In a conversation at CES 2026 with Jay Iyengar, CTO of Oshkosh, the company articulated a grounded, end-user-driven approach to AI that prioritizes real outcomes over experimentation and trust over novelty.
This approach offers a valuable lens into how AI adoption is evolving across industrial, municipal, defense, and infrastructure environments, where safety, reliability, and human impact matter more than technical sophistication.
Starting With the End User, Not the Algorithm
A defining characteristic of Oshkosh’s AI strategy is its focus on starting with the customer and end user, rather than the technology itself. Rather than developing AI in isolation and later searching for use cases, Oshkosh works forward from clearly defined problems faced by the people who operate, maintain, and rely on its vehicles and equipment every day.
These end users, firefighters, refuse collectors, airport ground crews, construction workers, and defense personnel operate in physically demanding, safety-critical environments. For them, AI must deliver measurable improvements in productivity, efficiency, and safety, not abstract technical promise. This philosophy fundamentally reshapes how AI is designed, deployed, and evaluated. Also, what is important for the AI and how it is deployed.
Bringing AI Into the Physical World
Unlike many enterprise AI initiatives that remain cloud-centric, Oshkosh is embedding AI directly into vehicles, machines, and robotic platforms through edge computing. This “physical AI” enables real-time perception and decision-making in environments where latency, connectivity constraints, and reliability requirements make cloud-only approaches impractical.
Examples highlighted at CES include autonomous refuse trucks that detect contamination, perimeter-security robots for airports, and robotic systems originally developed for defense training that are now being adapted for civilian infrastructure. By transferring mature technologies across domains, Oshkosh accelerates deployment while reducing risk, an increasingly important advantage as organizations look to operationalize AI rather than pilot it indefinitely.
For end users, this translates into AI that operates seamlessly in the background, assisting with complex or repetitive tasks without disrupting established workflows.
Augmentation Over Full Autonomy
A critical insight from the discussion was Oshkosh’s deliberate avoidance of “autonomy everywhere.” While the underlying technology may support fully autonomous operation, the company emphasizes what it calls “moments of autonomy,” selectively deploying automation where it adds clear value.
This distinction is essential in physical environments where trust and accountability are paramount. Rather than replacing humans, AI is used to augment their capabilities: reducing fatigue, improving consistency, and allowing skilled workers to focus on higher-value activities. From an adoption standpoint, this hybrid human-AI model is far more likely to gain acceptance among frontline users than fully autonomous systems that remove human control entirely.
Trust, Privacy, and Responsible AI Deployment
Trust emerged as a central theme in Oshkosh’s AI strategy. In physical AI systems, trust is not theoretical; it directly affects safety, adoption, and regulatory acceptance. Oshkosh treats trust as a system-level design principle spanning multiple layers.
This begins with carefully curated training datasets and supervisory AI models, continues with thoughtful decisions about where and how autonomy is deployed, and extends to privacy protections such as facial blurring and strict data retention policies. Importantly, Oshkosh focuses on retaining outcomes rather than raw data whenever possible, minimizing risk while still delivering operational insight.
Cybersecurity is treated as a baseline requirement, reinforcing that AI-enabled vehicles must be as secure as they are intelligent.
Customer-Driven Use Cases and Measurable Outcomes
What distinguishes Oshkosh’s approach is how directly customer demand shapes its AI roadmap. Municipal fleets are seeking more advanced detection of hazardous materials, such as batteries, in waste streams. Airports are requesting AI-enabled perimeter security and foreign object detection. Construction partners want task-specific robotics that address defined workflows rather than broad, generalized automation.
Across these use cases, customers consistently prioritize outcomes: time savings, labor efficiency, injury reduction, faster training, and improved talent attraction in industries facing persistent workforce shortages. AI becomes a tool not just for operational optimization, but for making physically demanding jobs safer, more sustainable, and more attractive to future workers.
Analyst Angle
From an analyst perspective, Oshkosh’s CES 2026 narrative underscores a broader shift in enterprise AI adoption. As AI moves from digital abstractions into physical systems, success will increasingly be defined at the point of use, by the driver in the cab, the worker on the tarmac, or the operator on the job site.
AI that augments human capability, earns trust through responsible design, and delivers measurable outcomes will define the next phase of digital transformation. Oshkosh’s end-user-first approach provides a compelling blueprint for how industrial organizations can move AI out of the lab and into real-world impact, where it ultimately matters most.
We see the approach Oshkosh and others are using as a way to directly connect to the value that will drive ROI and, ultimately, company value from AI.
Feel free to reach out and stay connected through robs@siliconangle.com, read @realstrech on x.com, and comment on our LinkedIn posts.
